In the proposed research, we plan to develop methods that examine the humoral response to prostate tumor antigens as a diagnostic and prognostic serum biomarker of disease. The method involves the use of protein microarrays produced by the two-dimensional liquid phase fractionation of cell lysates. A pl-based chromatographic separation is utilized in the first dimension to fractionate large numbers of proteins into discrete pH fractions over an extended pH range. In the second dimension, nonporous (NPS) reversed phase (RP) HPLC of each pH fraction will be used to separate out purified proteins to be spotted from the column eluent onto nitrocellulose slides. The result will be a protein microarray consisting of >2500 protein spots that can be tested against patient sera to examine resulting humoral response. The prostate cancer protein microarrays will be used to interrogate serum from healthy control individuals, patients with benign prostatic hyperplasia (BPH), and those with biopsy-proven prostate cancer. Using this approach, we hope to characterize a signature humoral response for patients with prostate cancer. The method arrays proteins from actual tumor cells so that antibodies in patient plasma may react to modified forms of a protein not necessarily produced by other methods. Proteins eliciting a humoral response to serum can be analyzed directly from liquid fractions by mass spectrometry where an accurate MW value and analysis of enzymatic peptide maps by CE-TOF-MS and MALDI-TOF-MS can be used for identification and determination of present modifications. Modified proteins may provide selective response to antibodies in sera that unmodified proteins may not provide. The presence of proteins in the liquid phase makes this method amenable to automated screening of large numbers of samples. The antibody response will be explored for its utility in screening/diagnosis (i.e. to supplement PSA testing) and prognosis (i.e., distinguish between indolent and aggressive prostate cancer). In addition, bioinformatic methods will be used to determine whether subclasses of prostate cancer exist that better discriminate among clinical outcomes. A predictive model will be developed cataloguing clinically meaningful markers based on humoral expression profiles from localized prostate cancer samples; thus, candidate tumor antigens can be identified for further experimental study. Candidate tumor antigens will be further validated using tissue microarrays.